Back in 1963, a statistician named Mollie Orshansky sat at her desk in the Social Security Administration and cobbled together a method that would quietly steer American social policy for decades. She took the cost of a bare-bones food plan, tripled it—because a 1955 survey said families spent about a third of their income on food—and called the result the federal poverty threshold. Six decades later, we’re still leaning on a metric built from assumptions that have long since collapsed under the weight of economic change.
As someone who has spent forty years studying the tangle of measurement and morality in public policy, I find this stubbornness intellectually sloppy and ethically indefensible. How we count the poor decides who gets help, how programs are funded, and ultimately how a society understands its obligations to its most vulnerable members. Get the counting wrong, and those obligations are quietly betrayed.
The Architecture of an Obsolete Metric
To see why the current poverty measure fails, you have to look at how it was built. Orshansky’s thresholds were never meant to capture the full cost of living. They were a rough shortcut, designed for a moment when food dominated household budgets and the safety net was far simpler than it is now.
The original calculation grabbed the Department of Agriculture’s “economy food plan”—the cheapest nutritionally adequate diet around—and multiplied its cost by three. That multiplier came from a 1955 survey showing families of three or more spent roughly a third of their after-tax income on food. Smaller households got slightly different multipliers, but the logic stayed the same: food spending served as the anchor for everything else.
Since then, the thresholds have been updated each year using just the Consumer Price Index for All Urban Consumers. So the poverty line tracks general price changes, not the real cost of housing, health care, child care, or transportation—all of which have climbed far faster than food and now swallow much bigger chunks of household budgets. The result is a threshold that drifts further from material reality every year.
The Disappearing Weight of Food
When Orshansky drew up her measure, American households put about 30 percent of their spending toward food. Today, that number sits around 10 percent, according to the Bureau of Labor Statistics. Meanwhile, housing costs have swelled from less than 20 percent of spending to over 30 percent, and health care has more than doubled its share of the average budget.
Pegging the poverty line to food made a certain sense when food was the biggest expense for poor families. It makes no sense now. A family that pours 50 percent of its income into rent and 10 percent into food simply isn’t captured by a threshold that assumes the opposite. The metric systematically undercounts the cost of surviving in an economy where shelter, medical care, and getting to work are far heavier burdens.
What the Official Measure Conceals
The fallout from this mismeasurement isn’t just academic. It shapes the daily lives of millions of Americans who slip through the gaps in eligibility formulas, funding allocations, and public awareness. When the official poverty line is set too low, people who can’t afford basic necessities are told, by the arithmetic of the state, that they aren’t poor.
Take housing. The official measure makes no geographic adjustment for cost-of-living differences. A family in rural Mississippi and a family in San Francisco get measured against the same dollar threshold, even though housing costs might differ by a factor of three or four. This isn’t a minor oversight; it’s a fundamental distortion that makes the statistic nearly useless for understanding regional hardship.
Medical expenses are a similar mess. The official measure counts pre-tax income, ignoring the heavy bite of payroll taxes and work-related costs. It also leaves out non-cash benefits like food assistance or housing vouchers from income, while simultaneously ignoring out-of-pocket medical costs that can push a family deep into deprivation even when their cash income looks okay on paper.
The Working Poor and the Fading Threshold
A single parent working full-time at the federal minimum wage pulls in about $15,000 a year. The 2024 poverty threshold for a family of two is roughly $19,720. By the official measure, that parent and child are poor. But bump the wage a little above the minimum, toss in some overtime, and the family crosses the line into statistical non-poverty—even if rent eats half their income and they rely on food pantries to eat. The threshold creates a cliff: small cash increases can trigger the loss of benefits far more valuable than the extra earnings, while the family’s material circumstances barely budge.
This dynamic shows up clearly in the research. Scholars at the University of Wisconsin’s Institute for Research on Poverty have documented how the official measure paints a distorted picture of economic well-being, especially for working families who receive the Earned Income Tax Credit and other refundable credits. Those credits get counted as income but often arrive as a lump sum after the year ends. The timing of resources matters enormously for families living on the edge, yet the annual income accounting of the poverty measure ignores it completely.
The Supplemental Measure: A Partial Correction
In 2010, the Census Bureau and the Bureau of Labor Statistics introduced the Supplemental Poverty Measure, a long-overdue effort to patch the worst holes in the official threshold. The SPM makes several real improvements: it adjusts thresholds for geographic housing cost differences, counts non-cash benefits and tax credits as income, and deducts necessary expenses like child care, medical out-of-pocket costs, and payroll taxes from resources. Its threshold is derived from actual spending on a broader bundle of necessities—food, clothing, shelter, and utilities—not from an outdated food multiplier.
The SPM reveals a poverty landscape that looks quite different from the official picture. Under the SPM, poverty is generally higher for the elderly—because medical expenses get subtracted from income—and lower for children, since benefits like nutrition assistance and the Earned Income Tax Credit are counted. The geographic adjustments shift the distribution of poverty toward high-cost states, particularly on the West Coast and in the Northeast.
But the SPM isn’t the official statistic used to determine eligibility for most programs. That designation still belongs to the Orshansky-based measure, which governs allocations for Medicaid, the Children’s Health Insurance Program, and a slew of other federal and state efforts. We have a better measure in hand, but we’ve declined to use it where it matters most.
The Political Economy of Threshold Maintenance
Why does an obviously flawed measure hang on? Partly bureaucratic inertia, partly the political fallout of updating the count. A more accurate poverty measure would, in many scenarios, produce a higher poverty rate. That higher rate would create pressure for more spending on anti-poverty programs—pressure that powerful interests would rather avoid. Keeping the threshold low acts as a quiet form of fiscal containment, trimming the number of people who appear to need help and limiting the apparent scope of the problem.
There’s a subtler dynamic at work, too. The official measure, by failing to count non-cash benefits, makes anti-poverty programs look less effective than they are. When food assistance and housing vouchers are invisible in the income calculation, the poverty rate seems stubbornly immobile, feeding narratives that government intervention is futile. Yet when the SPM counts those benefits, the poverty rate drops substantially, showing that programs do reduce hardship. The official measure thus serves a dual political function: it minimizes the apparent extent of poverty while simultaneously undercutting the case for the very programs that alleviate it.
Moral Dimensions of Mismeasurement
Behind the technical squabbles over multipliers and price indices sits a deeper moral question: What does a society owe its members, and how does it know when those obligations are being met? Measurement is never purely technical. It embodies normative choices about what counts as a minimally adequate life, which resources are considered essential, and whose suffering rises to the level of public concern.
A poverty threshold that ignores the cost of housing in a country where eviction can lead to homelessness, family separation, and cascading crises isn’t just inaccurate—it’s morally unserious. A measure that overlooks medical debt in a nation where illness is a leading cause of bankruptcy abdicates its responsibility to describe reality in a way that prompts action. Sticking with an obsolete measure, when a better one exists, is a choice with ethical weight.
I’ve argued elsewhere that poverty measurement should be guided by a capabilities approach, rooted in the work of Amartya Sen and Martha Nussbaum. The question isn’t merely whether a family’s income tops an arbitrary line, but whether they have the real freedom to achieve basic functionings: adequate nutrition, safe shelter, access to medical care, the ability to appear in public without shame. A threshold derived from 1960s food budgets cannot possibly answer that question for a 21st-century family.
Children and the Long Shadow of Statistical Neglect
The heaviest consequences of mismeasurement fall on children. Child poverty isn’t just a current deprivation; it predicts lower educational attainment, worse health, and diminished lifetime earnings. When the official measure undercounts child poverty, it understates the urgency of intervention during the years when intervention matters most. Research from the National Academy of Sciences has shown that a half-century of social science evidence on poverty’s effects on child development has been systematically disconnected from the statistical infrastructure used to allocate resources.
The SPM, by counting benefits that disproportionately help families with children, shows a lower child poverty rate than the official measure in some years. But that shouldn’t breed complacency. The SPM threshold itself is pegged to spending patterns at the 33rd percentile of the distribution—a choice that embeds a particular normative judgment about what counts as minimally adequate. That judgment deserves constant scrutiny, not passive acceptance.
Toward a More Honest Accounting
Reforming poverty measurement takes more than adopting the SPM as the official statistic, though that would be a big step forward. It requires a broader conversation about what we mean by poverty and what we’re willing to do about it. A few principles should guide that conversation.
First, thresholds should reflect the actual costs of meeting basic needs in specific places. The SPM’s geographic adjustments are a start, but they’re based on broad regional housing cost differences and don’t catch within-state variation. A family in Manhattan faces wildly different costs than a family in Buffalo, yet both get adjusted by the same New York state factor. More granular measures, like the Economic Policy Institute’s Family Budget Calculator, show it’s feasible to build thresholds that vary by county or metro area.
Second, the resource definition should account for the full range of expenses necessary for work and basic functioning. Child care, transportation to work, and out-of-pocket medical costs aren’t optional; they’re prerequisites for economic participation. Subtracting them from income before comparing against the threshold is a methodological necessity, not a statistical luxury.
Third, we might consider adopting multiple poverty measures for different purposes, rather than hunting for one all-purpose statistic. A measure designed to track long-term trends could use a quasi-relative threshold anchored to median consumption. A measure for program eligibility might use an absolute threshold based on the cost of a specific basket of goods. A measure to assess deprivation could fold in direct indicators of material hardship—food insecurity, housing instability, unmet medical needs—alongside income. No single number can serve all these purposes well.
FAQ
Why was the original poverty threshold based on food costs?
The original threshold, developed by Mollie Orshansky in 1963, used food costs as an anchor because a 1955 survey showed families spent about a third of their income on food. Orshansky took the cheapest nutritionally adequate food plan from the Department of Agriculture and tripled its cost to get a poverty line. At the time, food was the single biggest household expense, so it made a rough kind of sense as a proxy for overall need. The trouble is, that relationship no longer holds—housing, health care, and other costs have risen much faster than food prices.
How does the Supplemental Poverty Measure differ from the official measure?
The Supplemental Poverty Measure, introduced in 2010, makes several key fixes. Its threshold rests on spending for food, clothing, shelter, and utilities rather than a food multiplier. It adjusts for geographic differences in housing costs, counts non-cash benefits like food assistance and housing vouchers as income, and deducts necessary expenses such as child care, medical out-of-pocket costs, and payroll taxes from resources. These changes yield a more accurate picture of economic need, especially for groups like the elderly and families with children whose circumstances get distorted by the official measure.
Why hasn’t the U.S. adopted a more accurate poverty measure for program eligibility?
The official measure sticks around partly because of bureaucratic inertia—many programs have statutory formulas tied to the Orshansky thresholds—and partly because of political calculations. A more accurate measure would likely show higher poverty rates in many areas, creating pressure to boost anti-poverty spending. On top of that, the official measure’s failure to count non-cash benefits makes government programs look less effective, which suits certain political narratives. Changing the measure would take legislative action and a willingness to face what a more honest poverty count would reveal.
Does the official poverty measure affect real people’s lives?
Absolutely. The official poverty thresholds determine eligibility for a host of federal and state programs, including Medicaid, the Children’s Health Insurance Program, and subsidized school meals. When the threshold is set too low, families who can’t afford basic necessities get ruled ineligible for help. The measure also shapes public understanding of poverty’s scope and severity, influencing policy debates and funding priorities. Statistical choices made in Washington directly affect whether a struggling family gets assistance or slips through the cracks.